A High-Fidelity and Fast Modeling Method and System for the Assembly Surface of a Motor Iron Core Stamping Die

Abstract

The invention discloses a high-fidelity rapid modeling method and system for the assembly surface of a motor core stamping die; the method is as follows: First, establish a simulation ideal model of a concave template and an insert, and generate a mesh; collect the point cloud data of the assembly surface of the concave template and the insert. Second, respectively fit and establish a curved surface according to the point cloud data of the assembly surface of the concave template and the insert. Third, offset the surface mesh nodes of the simulation ideal model of the concave template and the insert to the corresponding curved surfaces established in the second step, and correspondingly offset the internal mesh nodes of the model to obtain an actual assembly model. The invention offsets the surface mesh nodes of the simulation model to the NURBS precise contact surface generated by the point cloud data, fully considers the geometric error of the assembly surface, makes the simulation closer to the actual situation, can accurately predict the deformation of the hole center after the die assembly, and provides an effective guidance for the die assembly.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital twin modeling, and particularly relates to a method and system for high-fidelity and rapid modeling of the assembly surface of a motor core stamping die. Technical Background

[0002] In the field of motor core stamping die assembly, the influence of geometric errors on the assembly surface on the assembly accuracy is universal. Even if dimensional tolerances and fit requirements are considered in the machining design stage, due to various factors such as machine tool accuracy, tool wear, geometric errors of the process system, and thermal and mechanical deformation, the contact surfaces of die inserts and templates will present irregular geometric shapes. The irregularity of this surface has a significant impact on the die assembly accuracy. Traditional simulation methods rely on ideal geometric surface modeling and ignore machining errors, resulting in a significant deviation between simulation predictions and actual assembly. In the traditional forward modeling process, the complex path of performing data processing, precise modeling, virtual assembly across platforms and then importing into a finite element simulation platform for multi-physics field analysis leads to problems such as low model reconstruction efficiency and easy distortion of multi-software data conversion. Due to the limitations of geometric model format conversion between different software, especially the differences in the discretized expression of complex surfaces, it is difficult to fit the actual geometric features during mesh generation.

[0003] To overcome the above problems, the present invention proposes a method for high-fidelity and rapid modeling using an innovative node offset method based on the geometric errors of the assembly surface of a motor core stamping die. Through a geometric-simulation integrated modeling framework, a full-chain closed-loop of error characterization, mesh optimization, and contact analysis is realized in a single CAE environment.

[0004] This method uses a coordinate measuring machine to scan the die surface to obtain point cloud data. In the finite element simulation platform, according to the drawings of the concave template and the insert, a three-dimensional model is established from the two-dimensional model, and then the point cloud data is subjected to error elimination and data filtering. Based on these surveying data, a NURBS surface model (Non-Uniform Rational B-Spline) is constructed, and the mapping of the nodes on the assembly surface to the NURBS surface is found. The mesh nodes of the holes in the template and the outer ring of the insert are offset to the corresponding NURBS surface. By flexibly adjusting the control points and weights, NURBS can accurately fit surfaces with irregular features such as local protrusions and depressions. In addition, the NURBS surface allows the analysis of geometric errors at different scales, taking into account both the global shape and local details. Through the construction of the geometric error surface and node offset, accurate surface modeling can be directly carried out in the simulation software, reducing errors in actual assembly, improving the overall accuracy of the die, and thus enhancing the assembly quality. This method can also be extended to precision scenarios such as gear pair meshing and bearing raceway assembly, but this patent focuses on the assembly surface modeling of motor core stamping dies. Compared with the traditional forward modeling method, this solution reduces the data interface redundancy by 80% compared with the prior art, and the modeling cycle is compressed by 60%. Summary of the Invention

[0005] The object of the present invention is to provide a high-fidelity rapid modeling method and system for the assembly surface of a motor core stamping die, aiming at the influence of geometric errors formed by the processing of assembly components on the assembly accuracy during the mechanical assembly process of the motor core stamping die. In traditional forward modeling methods, it is difficult to convert data formats between different software for establishing accurate surface simulations using multiple software, and there are errors in model integration, etc. Based on a new node offset technology, the present invention directly constructs an accurate surface model in a finite element simulation platform, and accurately simulates the geometric errors between assembly components through specific steps such as constructing a NURBS surface model and node offset, thereby improving the assembly accuracy of the die. This method provides a theoretical support for the mechanical design and assembly of precision dies, and can accurately predict the assembly results, which is a general and efficient accurate modeling and simulation technology.

[0006] In a first aspect, the present invention provides a high-fidelity rapid modeling method for the assembly surface of a motor core stamping die, which includes the following steps:

[0007] Step 1: Establish a simulation ideal model of the concave template and the insert, and generate a mesh. The mesh nodes on the assembly surface adopt an adaptive mesh; collect the point cloud data of the actual assembly surfaces of the concave template and the insert.

[0008] Step 2: Respectively fit and establish surfaces according to the point cloud data of the assembly surfaces of the concave template and the insert.

[0009] Step 3: Offset all the surface mesh nodes of the simulation ideal model of the concave template and the insert to the corresponding surfaces established in Step 2, and correspondingly offset the internal mesh nodes of the model to obtain an actual assembly model.

[0010] Preferably, the process of establishing the surface is as follows: Use the point cloud data as the control points Q k,l ; Through the cumulative chord length method, obtain the node parameters corresponding to the measured data points in the point cloud data, and establish a curve node vector; Construct a NURBS surface; Generate control point interpolation points on the NURBS surface according to the control points Q k,l ; Extract the extreme point set on the surface of the NURBS surface.

[0011] Preferably, the process of offsetting the nodes on the model surface in Step 3 is as follows:

[0012] (1) For each adaptive mesh node on the surface of the simulation ideal model , find the closest control point interpolation point Q k,l,t on the surface as the candidate target point .

[0013] (2) Calculate the candidate target points corresponding to each adaptive mesh node respectively The distance to the extreme point closest to itself. Judge each candidate target point separately to see if the corresponding distance satisfies being less than the target point movement threshold; if it does, use the candidate target point closest extreme point as the final target point ; otherwise, use the candidate target point directly as the final target point .

[0014] (3) Transfer each grid node on the surface of the simulation ideal model to the corresponding final target point respectively .

[0015] Preferably, the value of the target point movement threshold is 1×10 -5 ~3×10 -5 .

[0016] Preferably, the offset vector of the internal grid nodes of the simulation ideal model is expressed as follows:

[0017]

[0018]

[0019] where , are the offset vectors of surface grid node i and internal grid node j respectively; is the weight coefficient of surface grid node i to internal grid node j; is the coefficient for controlling distance attenuation; , are the coordinates of surface grid node i and internal grid node j before offset respectively; L ref is the average element side length; J j is the Jacobian determinant of the element to which internal grid node j belongs; J ref is the Jacobian reference value; AR j is the aspect ratio of the element to which internal grid node j belongs, AR ref is the aspect ratio reference value; α and β are the exponents for controlling the weights of Jacobian and aspect ratio respectively.

[0020] Preferably, use the actual assembly model for interference fit simulation, and extract the die offset distance from the simulation results as the assembly deformation between the holes of the die plate and the inserts.

[0021] Preferably, the sampling point spacing of the point cloud data in step one is 1 mm to 2 m; the point cloud data is processed by filtering and error separation. Further preferably, the preset tolerance range in error separation is 5% to 7%

[0022] Preferably, the simulation ideal models of the concave template and the insert are both set to elastic deformation, and a surface-to-surface contact relationship is established, and the actual frictional force is simulated through the contact friction penalty model.

[0023] In a second aspect, the present invention provides a high-fidelity and rapid modeling system for the assembly surface of a motor core stamping die, which is used to execute the foregoing high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die; the modeling system includes a coordinate measuring instrument, a NURBS surface generation module, and a finite element simulation module. The coordinate measuring instrument is used to collect point cloud data; the NURBS surface generation module is used to generate a surface according to the point cloud data; the finite element simulation module is used to establish a simulation model and offset the mesh nodes of the simulation model according to the surface generated by the point cloud data.

[0024] In a third aspect, the present invention provides a method for assembling a motor core stamping die, which includes the following steps:

[0025] Step 1: Execute the foregoing high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die to model the die holes and inserts to be assembled.

[0026] Step 2: For the die hole model and the insert model obtained in Step 1, perform interference fit simulations at different relative angles to obtain the die offset distances at different relative angles; take the relative angle with the smallest die offset distance as the installation angle.

[0027] Step 3: Perform an interference connection on the actual die holes and inserts to be assembled according to the installation angle determined in Step 2.

[0028] The beneficial effects of the present invention are as follows:

[0029] 1. The present invention provides a new node offset method, which offsets the surface mesh nodes of the simulation model to the NURBS precise contact surface generated by the point cloud data, fully considering the geometric error of the assembly surface, making the simulation closer to the actual situation, and being able to accurately predict the deformation of the hole center after die assembly, providing effective guidance for die assembly.

[0030] 2. The present invention adopts the strategy of "first discretizing the basic mesh and then offsetting it to the NURBS precise surface", which significantly simplifies the mesh discretization process while retaining the original geometric accuracy. It not only avoids the compatibility problem of cross-platform mesh generation but also makes the mesh nodes discrete to the extreme points of the surface, fully ensuring the high-fidelity geometric error of the simulation model.

[0031] 3. The present invention realizes an accurate digital twin model containing machining errors that can be directly used for simulation analysis. Through a geometric-simulation integrated modeling architecture, a full-chain closed loop of error characterization, mesh optimization, and contact analysis is achieved in a single CAE environment, reducing the data interaction links between software, lowering the risks of data loss and error introduction, improving the modeling efficiency and simulation accuracy, and simplifying the simulation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The flowchart of the present invention.

[0033] Figure 2 The ideal simulation model diagram of the concave template and the insert obtained in step S200 of the present invention.

[0034] Figure 3 The actual assembly model diagram of the concave template and the insert obtained in step S400 of the present invention.

[0035] Figure 4 The contact pressure distribution diagram of the concave template and the insert after interference fitting of the die obtained in step S500 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] The following uses specific specific examples to illustrate the embodiments of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0037] Embodiment 1

[0038] As Figure 1 shown, a high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die includes the following steps:

[0039] Step S100 determines the machining drawing information and process requirements, and measures the point cloud data of the contact surface:

[0040] First, according to the actual machining requirements of the motor core die, obtain and confirm the die drawing information and the corresponding process requirements. This drawing information includes but is not limited to: the nominal diameter, tolerance range, shape error, position error, interference fit size, etc. of the concave template and the insert. This information provides basic data for subsequent measurement and simulation modeling.

[0041] Before assembly, use a coordinate measuring machine (CMM) to perform high-precision measurements on the assembled surfaces of the finish-machined concave template and insert. Before using the coordinate measuring machine, it is necessary to calibrate the measuring instrument with a standard ball according to the instrument operation procedures to ensure that the measurement error is within the allowable range. During the calibration process, measure the standard ball multiple times, obtain the measurement data, compare it with the standard value, and calculate the measurement deviation. If the deviation exceeds the specified range, the measuring instrument needs to be adjusted or repaired until the accuracy requirements are met.

[0042] S110 Clean the concave template and insert before measurement to ensure that there are no interfering factors such as oil stains and impurities on the surface.

[0043] S120 The coordinate measuring machine uses a sub-micron-level probe to ensure that the measurement accuracy can meet the design requirements. The length of the probe is adapted according to the depth of the measurement parts of the concave template and insert to ensure that the probe can accurately contact the measurement surface during the measurement process without excessive deformation or interference. At the same time, check the installation firmness and perpendicularity of the probe. After installation, use a standard gauge block to perform a simulated measurement to verify the measurement repeatability of the probe.

[0044] S130 Use a multi-point sampling method to measure the inner surface of the concave template, and densify the sampling points in the key mating areas (such as the hole wall and positioning surface that mate with the insert); in this embodiment, the sampling point spacing is set to 1 mm; in some embodiments, the sampling spacing in other areas is appropriately adjusted according to the surface curvature and size, but the minimum is not less than 2 mm. For the outer surface of the insert, key parts in contact with the concave template are also measured emphatically to ensure that the geometric shape information of the assembled surface can be obtained comprehensively and accurately.

[0045] The above steps use a coordinate measuring machine to record the point cloud data of the inner wall of the concave template and the outer surface of the insert, and these data will be used for subsequent geometric error analysis and simulation modeling.

[0046] Step S200 Finite element simulation modeling

[0047] Based on the drawing information of the motor core mold, use finite element simulation software to construct a simulation ideal model of the motor core mold assembly.

[0048] S210 Based on the drawing information of the motor core mold, use drafting software to draw a 2D sketch of the motor core mold and convert its format into the standard file IGES format for easy import into the simulation software to establish a 3D solid model. Check the integrity and accuracy of the geometric elements during the conversion process to ensure that there is no data loss or error when importing into the simulation software.

[0049] S220 Import the sketch into the simulation software, establish 3D solid models of the concave template and insert in the motor core mold, and perform the assembly.

[0050] S230 sets the material properties. The material of the concave template in the motor core mold is die steel DC53, and the material of the insert is cemented carbide ZNF26. Since the contact deformation between the concave template and the insert is an interference fit and their deformations are elastic deformations, it is necessary to set the material density, elastic modulus, and Poisson's ratio. In this embodiment, the material parameters of the concave template and the insert are shown in Table 1.

[0051] Table 1 Mold material parameters

[0052] Material <![CDATA[Density / (t / mm 3 )]]> Young's modulus / (MPa) Poisson's ratio DC53 <![CDATA[7.8×10 -9 > <![CDATA[2.31×10 6 > 0.285 ZNF26 <![CDATA[1.4×10 -8 > <![CDATA[5.86×10 6 > 0.235

[0053] S240 establishes a surface-to-surface contact relationship between the concave template and the insert, and enables the contact friction penalty model to simulate the frictional force in the actual contact process. By referring to relevant friction coefficient manuals and actual empirical data, and combining the working environment of the mold (such as lubrication conditions, working temperature, etc.), the friction coefficient is determined to be 0.15. The normal contact is set to hard contact to ensure that the contact surfaces do not penetrate each other during the assembly process.

[0054] S250 discretizes the mold and the insert into meshes. The mesh size is matched with the sampling point spacing in step S130. In this embodiment, the mesh size is taken as 1 mm. Based on the actual processing requirements, the corresponding concave template mesh and insert mesh in the motor core mold are selected as the adaptive mesh regions, and the mesh nodes on the inner surface of the concave template and the outer surface of the insert are selected as the adaptive mesh nodes to facilitate the establishment of the precise surfaces of the concave template and the insert in the subsequent simulation. An analysis step is established, and the output variables are set to stress, strain, displacement field, contact pressure, and node coordinates.

[0055] The concave template model and the insert model obtained in this step are as Figure 2 shown.

[0056] Step S300 processes the point cloud data, eliminates random errors, and extracts valid data.

[0057] In the specific implementation of the present invention, after obtaining the point cloud data collected by the coordinate measuring machine, it is necessary to preprocess the data because there may be random errors caused by measurement equipment or environmental conditions in the point cloud data. The specific process is as follows:

[0058] S310 data pre-screening: Based on the filtering algorithm of the least squares method, the obvious outliers in the measured point cloud data are identified, and the noise points generated during the measurement process are filtered out by setting a threshold.

[0059] S320 error separation: Based on the curve fitting method of the least squares method and the residual filtering method, the point cloud data is fitted into a preliminary surface model, and then by calculating the deviation of each measurement point from the fitted surface, the points exceeding the preset tolerance range are eliminated.

[0060] S330 Extract valid data: After removing random errors, further analyze the remaining data and extract the data point Q of the value points suitable for establishing an accurate surface k ,l (k = 0, 1, …, n; l = 0, 1, …, m), where n + 1 and m + 1 are the number of layers of point cloud acquisition and the number of points in each layer respectively. These valid data should cover the entire contact surface and ensure that the point cloud density is high enough to meet the requirements of NURBS surface fitting.

[0061] Through the processing of the above steps, the obtained point cloud data not only removes the random errors in the measurement, but also extracts the accurate surface geometric information, ensuring the accuracy of the simulation modeling.

[0062] Step S400 NURBS surface equation development and knot offset

[0063] S410 Write a subroutine for the NURBS surface equation through secondary development to construct the NURBS surface model of the contact surface between the concave template and the insert. The specific steps are as follows:

[0064] Determine the degrees of the NURBS surface as p and q, and the weight factor of the value points Q k,l is w k,l = 1, for the value points Q of the point cloud data after error separation k ,l (k = 0, 1, …, n; l = 0, 1, …, m); n + 1 and m + 1 are the number of layers of point cloud acquisition and the number of points in each layer respectively; use the cumulative chord length method for parametric calculation to obtain the knot parameters corresponding to all measurement data points in the same layer , and further calculate the curve knot vector U as follows:

[0065] (1)

[0066] (2)

[0067] (3)

[0068] (4)

[0069] (5)

[0070] where are the coordinates of two adjacent measurement data points respectively; k and l are the serial numbers of the measurement data value points in the point cloud data, and d is the sum of the distances between two adjacent measurement data points.

[0071] Similarly, calculate the knot parameters and the curve knot vector V as follows:

[0072] (6)

[0073] (7)

[0074] (8)

[0075] (9)

[0076] (10)

[0077] Using the NURBS surface reconstruction technology, based on the NURBS surface equation, construct the NURBS surface, whose surface must pass through the control points, and obtain the control point P through the following formula k,l as follows:

[0078]

[0079] wherein, represents the rational basis function of the p-th order NURBS curve determined by the i-th data point; represents the rational basis function of the q-th order NURBS curve determined by the j-th data point; P k,l represents the vertex of the NURBS surface control grid. Calculate the control points, so as to obtain the interpolation points Q k,l,t , and obtain the set of extreme points of the distance from the axis in the NURBS surface (a total of TOTAL) POINT_MAX_MIN = (x t , y t , z t ), (t = 1, 2,..., TOTAL).

[0080] In this embodiment, a high-precision numerical calculation method (such as the Gaussian elimination method or the LU decomposition method) is used to ensure the accuracy of the control point calculation, and the calculation results are verified and analyzed. For example, it is checked whether the control points meet the geometric continuity and smoothness requirements of the curve.

[0081] In the process of constructing the surface in this embodiment, the shape and characteristics of the surface are analyzed and verified. For example, it is checked whether the curvature change of the surface conforms to the geometric characteristics of the actual assembly surface, and through the comparison and visualization display with the original point cloud data, it is ensured that the surface equation can accurately describe the geometric shape of the assembly surface.

[0082] S420 Find the mapping of the nearest point or extreme point of the nodes on the simulation ideal model on the NURBS surface, and offset the mesh nodes of the holes in the template and the outer ring of the insert to the corresponding NURBS surface. The specific steps are as follows:

[0083] For each grid node on the simulation ideal model , the nearest value interpolation point on the NURBS surface is obtained by using the optimization method .

[0084] After step S421, each adaptive grid node on the surface of each simulation ideal model searches for the nearest value interpolation point Q on the NURBS surface k,l,t , as the candidate target point ;

[0085] For the candidate target point corresponding to each adaptive grid node , calculate the distance between each candidate target point and the closest extreme point again. The calculation formula is as follows:

[0086]

[0087] Judge whether the condition that the distance deter_min is less than 10 -5 is satisfied. If it is satisfied, take the closest extreme point of the candidate target point as the final target point ; otherwise, take the candidate target point directly as the final target point ; finally, transfer each adaptive grid node on the surface of the simulation ideal model to the corresponding final target point on the NURBS surface respectively ; its corresponding displacement vector .

[0088] Move each adaptive grid node in turn according to its corresponding displacement vector ; after each adaptive grid node moves, the controlled movement of the internal grid nodes of the simulation ideal model is realized by any Arbitrary Lagrangian-Eulerian (ALE) method. The internal grid nodes (a total of total_inter) are represented by a set . Under the minimum grid distortion, the Jacobian and aspect ratio of the original grid will change. To ensure high grid quality, a weight coefficient w ij is proposed in this embodiment to give an offset correction amount to the remaining grid nodes to ensure that the quality of the grid meets the requirements. The remaining grid nodes (a total of total_other) refer to the nodes inside the model, that is .

[0089] To meet the smoothness principle of the ALE adaptive mesh, this embodiment proposes a grid-based Jacobian and aspect ratio coefficient for the correction of the remaining mesh nodes.

[0090]

[0091]

[0092] Where is the displacement vector of the nodes in other node sets, is the coefficient for controlling the distance attenuation, L ref is the average element side length, J j is the node is the Jacobian determinant of the element to which the node belongs, J ref is the Jacobian reference value, AR j is the node is the aspect ratio of the element to which the node belongs, AR ref is the aspect ratio reference value, is the exponent for controlling the weights of the Jacobian and aspect ratio.

[0093] Through the ALE adaptive mesh offset, the boundary nodes of the adaptive mesh will be offset to the exact surface. Based on the smoothness principle and the Jacobian and aspect ratio coefficients of the mesh, the remaining mesh nodes will also be adjusted accordingly to ensure the convergence of the mesh quality. After all the adaptive mesh nodes have completed their movement, after the remaining mesh nodes are corrected by the ALE adaptive mesh, a new mesh is redrawn to obtain the actual assembly model of the motor core stamping die.

[0094] The actual assembly model obtained in this step is as Figure 3 shown.

[0095] In step S500, an interference fit simulation is performed. According to the simulation results, the offset distance of the concave die is extracted and compared with the actual measurement results to verify the accuracy of the simulation model.

[0096] In S510, the constructed actual assembly model and the set parameters are submitted to the solver for simulation operations to obtain the displacement deformation, mesh stress, mesh strain, and contact pressure in the contact area (as Figure 4 shown). According to the displacement deformation at different positions, the offset of the center of the hole in the concave die plate is obtained as the offset distance of the concave die; the offset distance of the concave die is used as the assembly deformation between the hole in the template and the insert. Before submission, check the integrity of the model and the rationality of the parameters to ensure that the solver can correctly perform the simulation calculation. During the simulation calculation, monitor the calculation progress and resource occupancy in real time, such as the memory usage and CPU usage rate, and adjust the calculation parameters or hardware resource allocation according to the actual situation to ensure that the simulation calculation can be successfully completed.

[0097] S520 Extract simulation offset data: According to the simulation results, locate the node coordinates corresponding to the actual measurement positions (such as the edge nodes of the insert), and fit the offset of the concave template hole position by weighted least squares to ensure the weight distribution in the key areas to improve the fitting accuracy.

[0098] S530 Obtain actual offset data: Using the actual assembly data collected by the coordinate measuring machine, after error compensation, fit the actual center offset by the least squares method to eliminate the influence of measurement noise on the results.

[0099] S540 Comparative analysis and model optimization: Statistically analyze the simulation offset data and the actual offset data (such as percentage error, standard deviation, etc.). If the error is within the allowable range (≤5%), the model is valid and can be directly used for mold design and process optimization; if it exceeds the tolerance, it is necessary to iteratively correct the contact model parameters (such as friction coefficient, material properties) or re-evaluate the machining error data until the simulation results converge with the actual assembly error.

[0100] To verify the superiority of this method, take the motor core mold of an enterprise as an example. The inner diameter tolerance of its concave template is Φ35.5±0.005mm. The assembly point cloud data is obtained by the coordinate measuring machine, and the NURBS surface model is constructed by the node offset method (the control point density is 1 / 3 of the traditional method). The simulation results show that the predicted value of the insert center offset is 15.3μm in the horizontal direction and 13.2μm in the vertical direction. The absolute error with the actual assembly measurement value of 15.5μm in the horizontal offset is 0.2μm, and the relative error is only 1.3%. The actual assembly measurement value in the vertical direction is 13.5, and the relative error is only 2.2%. While the predicted value of the traditional ideal model is 12.5μm in the horizontal direction, and the relative error reaches 19.4%. In the vertical direction, it is 11.9, and the relative error reaches 11.8%.

[0101] Table 2 Comparison between the ideal model method and the method of the present invention

[0102] Index Traditional ideal model Method of the present invention Improvement effect Horizontal error of hole center offset prediction (μm) 3.0 (error 19.4%) 0.2 (error 1.3%) Accuracy improved by 74.4% Vertical error of horizontal error prediction (μm) 1.6 (error 11.8%) 0.3 (error 2.2%) Accuracy improved by 81.3% Surface detail loss rate 18% <2% Detail restoration improved by 88%

[0103] To illustrate the simplicity of this method, a motor core mold of an enterprise is taken as an example. The inner diameter tolerance of the concave template is Φ35.5±0.005mm. By using a coordinate measuring machine to obtain the point cloud data of the assembly points, compared with the traditional forward modeling process in the existing precise modeling technology that requires cross-platform execution of data processing, parametric modeling, reverse engineering modeling, virtual assembly, and then importing into the finite element simulation platform for multi-physical field analysis, and the method of the present invention directly establishes an accurate surface model by using the node offset method to construct a NURBS surface model in the finite element simulation platform. The simulation results show that the horizontal offset of the insert center in the method of the present invention is 15.3μm and the vertical offset is 13.2μm. The absolute error compared with the actual assembly measurement value of 15.5μm in the horizontal direction is 0.2μm, and the relative error is only 1.3%. The actual assembly measurement value in the vertical direction is 13.5, and the relative error is only 2.2%. The predicted value of the traditional forward modeling technology is 15.2μm in the horizontal direction, with a relative error of 1.4%, and 13.1μm in the vertical direction, with a relative error of only 2.9%. This method compresses the modeling cycle from 4 hours to 1.6 hours, and the efficiency is increased by 60%.

[0104] Table 3 Comparison between the traditional forward modeling method and the method of the present invention

[0105] Index Traditional forward modeling method Method of the present invention Improvement effect Data interaction step 5 times 1 time Reduced by 80% Modeling time (hours) 4.0 1.6 Shortened by 60% Grid division time (hours) 1.0 0.5 Shortened by 50% Horizontal error of offset prediction (μm) 0.3 (error 1.4%) 0.2 (error 1.3%) Accuracy is almost the same Vertical error of offset prediction (μm) 0.3 (error 2.2%) 0.4 (error 2.9%) Accuracy is almost the same Number of model correction iterations 5 times 2 times Reduced by 60%

[0106] The above data shows that while ensuring sub-micron accuracy, this method greatly simplifies the modeling process and provides technical support for enterprises to achieve "one-time modeling and accurate prediction".

[0107] This embodiment also provides a high-fidelity and rapid modeling system for the assembly surface of a motor core stamping die; this modeling system includes a coordinate measuring machine, a NURBS surface generation module, and a finite element simulation module. The coordinate measuring machine is used to collect point cloud data; the NURBS surface generation module is used to generate a surface according to the point cloud data; the finite element simulation module is used to establish a simulation model and offset the mesh nodes of the simulation model according to the surface generated by the point cloud data.

[0108] Embodiment 2

[0109] An assembly method for a motor core stamping die includes the following steps:

[0110] Step 1: Model the die holes and inserts to be assembled, and the modeling process is the same as that in Embodiment 1.

[0111] Step 2: For the die hole model and the insert model obtained in Step 1, perform interference fit simulations at different relative angles to obtain the die offset distances at different relative angles; take the relative angle with the smallest die offset distance as the installation angle.

[0112] In this embodiment, interference fit simulations were performed for three different relative angles; these three relative angles increased by 45° in sequence, and the die offset distances in the horizontal and vertical directions are shown in the following table:

[0113] Table 1 Offset Distance Table for Different Relative Angles

[0114] Rotation angle Horizontal die offset distance Vertical die offset distance 0° 14.9μm 12.8μm 45° 15.1μm 13.6μm 90° 15.6μm 12.8μm

[0115] The relative angle of 45 degrees with the smallest die offset distance was taken as the installation angle.

[0116] Step 3: Perform an interference connection on the actual die hole and the insert to be assembled according to the installation angle determined in Step 2. Through actual measurement, the horizontal offset measurement value of the insert center was 15.1 μm, and the vertical offset distance was 12.9. Compared with the simulation of a 45-degree counterclockwise rotation, the error in the horizontal direction was only 0.2 μm, and the error in the vertical direction was only 0.1 μm.

[0117] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as a limitation of the present invention itself. Various changes in form and detail may be made without departing from the spirit and scope of the present invention as defined by the appended claims.

Claims

1. A high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die, characterized in that: It includes the following steps: Step 1: Establish a simulation ideal model of the concave template and the insert, generate a mesh, and use adaptive meshes for the mesh nodes on the assembly surface; And collect the point cloud data of the assembly surfaces of the concave template and the insert; Step 2: Fit and establish surfaces based on the point cloud data of the concave template and the insert assembly surface respectively. The process of establishing the surface is as follows: Use the point cloud data as the profile points Q of the surface k,l ; Through the cumulative chord length method, obtain the node parameters corresponding to the measurement data points in the point cloud data, and establish the curve node vector; Construct the NURBS surface; According to the profile points Q k,l Generate profile interpolation points on the NURBS surface; Extract the set of extreme points on the surface of the NURBS surface; Step 3: Offset the surface mesh nodes of the simulation ideal model of the concave template and the insert to the corresponding surfaces established in Step 2, and correspondingly adjust the internal mesh nodes of the simulation ideal model to obtain the actual assembly model; The process of offsetting the nodes on the model surface in Step 3 is as follows: (1)For each adaptive grid node on the surface of the simulation ideal model , find the closest value interpolation point Q on the surface k,l,t as the candidate target point ; (2) Calculate the distances between the candidate target points corresponding to each adaptive grid node and the extreme points closest to themselves respectively; judge whether the distances corresponding to each candidate target point satisfy being less than the target point movement threshold respectively; if satisfied, use the extreme point closest to the candidate target point as the final target point ; otherwise, directly use the candidate target point as the final target point ; ​ (3) Transfer each grid node on the surface of the simulation ideal model to the corresponding final target point respectively .

2. A high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die according to claim 1, characterized in that: The value of the target point movement threshold is 1×10 -5 ~3×10 -5 .

3. A high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die according to claim 1, characterized in that: The offset vector of the internal grid nodes of the simulation ideal model has the following expression: ； ； Among them, and are the offset vectors of the surface grid node i and the internal grid node j, respectively; is the weight coefficient of the surface grid node i with respect to the internal grid node j; is the coefficient for controlling distance attenuation; and are the coordinates of the surface grid node i and the internal grid node j before offset, respectively; L ref is the average element side length; J j is the Jacobian determinant of the element to which the internal grid node j belongs; J ref is the Jacobian reference value; AR j is the aspect ratio of the element to which the internal grid node j belongs, AR ref is the aspect ratio reference value; α and β are the exponents for controlling the weights of the Jacobian and the aspect ratio, respectively.

4. A high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die according to claim 1, characterized in that: Use the actual assembly model for interference fit simulation, and extract the die offset distance from the simulation results as the assembly deformation between the hole of the concave template and the insert.

5. A high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die according to claim 1, characterized in that: In Step 1, the sampling point spacing of the point cloud data is 1 mm to 2 m; the point cloud data is processed by filtering and error separation.

6. A high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die according to claim 1, characterized in that: The simulation ideal models of the concave template and the insert are both set to elastic deformation, and a surface-to-surface contact relationship is established, and the actual frictional force is simulated through the contact friction penalty model.

7. A high-fidelity and rapid modeling system for the assembly surface of a motor iron core stamping die, characterized in that: It is used to execute a high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die as described in Claim 1; this modeling system includes a coordinate measuring instrument, a NURBS surface generation module, and a finite element simulation module; the coordinate measuring instrument is used to collect point cloud data; the NURBS surface generation module is used to generate a surface according to the point cloud data; the finite element simulation module is used to establish a simulation model and offset the mesh nodes of the simulation model according to the surface generated from the point cloud data.

8. An assembly method for a motor iron core stamping die, characterized in that: It includes the following steps: Step 1: Execute a high-fidelity and rapid modeling method for the assembly surface of a motor core stamping die as described in Claim 1 to model the die hole and the insert to be assembled, Step 2: For the die hole model and the insert model obtained in Step 1, perform interference fit simulations at different relative angles to obtain the die offset distances at different relative angles; take the relative angle with the smallest die offset distance as the installation angle; Step 3: Perform an interference connection on the actual die hole and the insert to be assembled according to the installation angle determined in Step 2.

Images